This adapter provides a shim for Docker Engine that lets you control Docker via the Kubernetes Container Runtime Interface.
Take a look at the official docs for more information.
For users running 0.2.5
or above, the default network plugin is cni
. Kubernetes 1.24+ has removed kubenet
and
other network plumbing from upstream as part of the dockershim
removal/deprecation. In order for a cluster to become
operational, Calico, Flannel, Weave, or another CNI should be used.
For CI workflows, basic functionality can be provided via containernetworking/plugins
.
Mirantis and Docker have agreed to partner to maintain the shim code standalone outside Kubernetes, as a conformant CRI interface for the Docker Engine API. For Mirantis customers, that means that Docker Engine’s commercially supported version, Mirantis Container Runtime (MCR), will be CRI compliant. This means that you can continue to build Kubernetes based on the Docker Engine as before, just switching from the built in dockershim to the external one.
Mirantis and Docker intend to work together on making sure it continues to work as well as before and that it passes all the conformance tests and continues to work just like the built in version did. Mirantis will be using this in Mirantis Kubernetes Engine, and Docker will continue to ship this shim in Docker Desktop.
You can find more information about the context for this tool in Don't Panic: Kubernetes and Docker and on the Mirantis blog.
We can be found on the Kubernetes Slack in the #cri-dockerd
channel.
The easiest way to install cri-dockerd
is to use one of the pre-built binaries or
packages from the releases page.
There are numerous supported platforms and using a pre-built package will install
the binary and setup your system to run it as a service.
Please refer to your platform's documentation for how to install a package for additional help with these.
Note: the release packages will install to /usr/bin which is reserved for binaries managed by a package manager. Manual installation doesn't involve a package manager and thus uses /usr/local/bin and the service file must be edited to reflect this.
If you would like to install the project manually, you will need to place the binary
somewhere in your PATH
and setup a service to run it. The following command is
a manual install for a Linux system using systemd:
git clone https://github.com/Mirantis/cri-dockerd.git
The above step creates a local directory called cri-dockerd
which you will need for the following steps.
To build this code (in a POSIX environment):
cd cri-dockerd
make cri-dockerd
To build for a specific architecture, add ARCH=
as an argument, where ARCH
is a known build target for golang
You can find pre-compiled binaries and deb/rpm packages under:
https://github.com/Mirantis/cri-dockerd/releases
Where VERSION
is the latest available cri-dockerd version:
https://github.com/Mirantis/cri-dockerd/releases/download/v${VERSION}/cri-dockerd-${VERSION}.${ARCH}.tgz
To install, on a Linux system that uses systemd, and already has Docker Engine installed
# Run these commands as root
cd cri-dockerd
mkdir -p /usr/local/bin
install -o root -g root -m 0755 cri-dockerd /usr/local/bin/cri-dockerd
install packaging/systemd/* /etc/systemd/system
sed -i -e 's,/usr/bin/cri-dockerd,/usr/local/bin/cri-dockerd,' /etc/systemd/system/cri-docker.service
systemctl daemon-reload
systemctl enable --now cri-docker.socket
The default network plugin for cri-dockerd
is set to cni
on Linux. There are
a few ways to change this depending on how you are running the binary.
--network-plugin=${plugin}
can be passed in as a command line argument when
- running the binary directly
- adding to
/usr/lib/systemd/system/cri-docker.service
if a service isn't enabled - adding to
/etc/systemd/system/multi-user.target.wants/cri-docker.service
if a service is enabled
Run systemctl daemon-reload
to restart the service if it was already running.
If you would like to build the project yourself, you will need to have Go installed. You can find directions for installing the latest version on its website:
Install the latest version of Go
Once you have Go installed, you can build the project by running the following command:
make cri-dockerd
This will output the binary to the project's root directory as cri-dockerd
.
You can then run it directly or install it using the manual process above.
To build for a specific architecture, add ARCH=
as an argument, where ARCH
is a known build target for Go.
ARCH=amd64 make cri-dockerd
When developing, it is nice to have a separate environment to test in so that
you don't have to worry about breaking your system. An easy way to do this is
by setting up a minikube cluster since it uses cri-dockerd
by default. Follow
the minikube installation instructions
to get it installed.
You'll then be able to create a cluster in minikube's VM by running:
minikube start
Once the cluster is up, we have a make
command that will build cri-dockerd
and swap it out for the version running in the cluster. You can run this command
by running:
make dev
This folder contains the files used to generate the cri-dockerd
documentation.
The docs are generated using Hugo and the Geekdocs theme.
The docs can be ran locally with hot-reloading to make editing easier. To do so, run the following command in the project's root directory:
make docs
This will launch the development server that is included with Hugo. You can then access the docs at http://localhost:1313/